Performance analysis of a hybrid bike sharing system: A service-level-based approach under censored demand observations

We investigate a hybrid bike-sharing system. We carry out a usage pattern and demand analysis on the booking data of the system and include the effects of censored demand in a service level analysis. Service levels are used as meaningful measures for evaluating the customer-oriented performance of bike-sharing systems. Our results show that service levels are overestimated when ignoring unobservable, censored demand effects. Furthermore, there are significant differences between free-standing and station-based bikes. Based on these results, an adjusted incentive and reposition policy could increase the booking number of free-standing bikes and thus customer satisfaction as well as the system's profitability.